pbem hosting hell!
 

Hmmmm, epic pbem game conclusion, large fleet battles, thousands of fighters ...

Just how long could it take the hosting program to produce a set of turn files from a complex endgame battle across many systems with thousands of involved ships?

I ask because I have just returned from setting SEIV to do the hosting (taking 2 hours to watch a DVD) and still returning to find the progress bar on 72%!

The program appeared to be working still - the mouse moved, the hd light kept flickering but the progress bar did not appear to have moved from where I left it.

Could this take days?

Personally, I suspect the fighter defence stack bug which could create multiple redundant pointless combats with fighter Groups unable to hit each other.

Any hints on how to speed up hosting?

Any word on the patch (to resolve the fighter issues)?

Ah well, leaving my laptop to labour on overnight to hope the hosting will be done in the morning....

Re: pbem hosting hell!
 

Perhaps u should try Play-by-Web?

it makes the turn processing so much easier, particularly with a larger number of players spread out over many timezones.

At the moment it only works with Version 1.30, so that doesnt help u. But later Versions should work with the system, and the fighter problem should be resolved by then.
http://seiv.pbw.cc

Re: pbem hosting hell!
 

We did one final battle, we had decided that one player had won already. We got 5 fleets of different empires together and let them fight it out.
There was 50+ in three empires.

It took along time to do the roll over, not as long as you though.

The bad news is that I could not watch the reply as it would say that it had an addressing problem.

Re: pbem hosting hell!
 

The replay aint worth much anyway since you don't see damage taken by the ships. It would also be much better if you could step from ship to ship in every turn and see exactly what it does and not as it now with ship after ship going at it with no control by you until the turn ends. And a better combat report in the log wouldn't hurt either. A table with ships killed, captured and damaged sortable by size and design would be good.

Re: pbem hosting hell!
 

.... well, the host program is still running ... 8 hours and counting - the computer still responds, the mouse moves, the hard-drive whirrs, but the message still shows 72% complete!

Re: pbem hosting hell!
 

The task list in windows me does indeed show the program as not responding - but then, it shows that even when processing a turn which does later work. In this case I have tried selectively ignoring player turns and with a couple removed the hosting process does end after around 40mins (in the interim the symtoms are as for whole hosting process). Running the turn on an NT server with tonnes of memory shows the SEIV task at 100% usage at 72% of the process.

If the host program can't be run at this point then what? Abandoning a pbem game after 102 turns because of the processing overload is a bit of a bitter pill to swallow.

I suspect that fighters are the cause .... hundreds of mini-stacks fighting other mini stacks with 1% hit rates and multiple redundant combats ... would explain the length of time it is taking.

Does anyone have a reason to suppose the host program is unstable and wouldn't be able to continue working after 8+ hours btw?

Regards

Re: pbem hosting hell!
 

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>If you check the task list it will probably tell you Se4 is not responding<HR></BLOCKQUOTE>WARNING: a "not responding" tag can very easily be a false positive.

"Not responding" will show up if your application does not draw to the screen for 60 seconds or so. If you time it right, you can see NR tags when SE4 is between turns (easier to find when using a slow machine).

Re: pbem hosting hell!
 

like i said, the "not responding" thing is pretty common whil e processing my current game. At some stage in the process the completed percentage always waits a few minutes (presumably while combat or something is resolved) - I am assuming that in this case the (now) 11 hour freeze on 72% is actually doing something ... there is certainly hd access still ... question is, how long could this take theoretically?

Something I guess only Aaron could really take a stab at?

Re: pbem hosting hell!
 

Wow thats a long time. It would be a shame if it wouldn't process, sort of like reading a good book and finding the Last chapter is missing.

Re: pbem hosting hell!
 

Prhaps you could upload the file and let someone els (not me!!!) with a faster (or at least different) machine have a stab at it - unlss there are scurity issues, it a vs human game.


--
Space Vikings! http://www.sandman43.fsnet.co.uk/viking.htm

Re: pbem hosting hell!
 

we have a sort of workaround so the game will be able to continue ... it involves one of the players not using fighters in the way he wanted to though

*looks like largescale fighter combat launched from carriers is unviable with the current patch*

... shouldn't really effect the strategic dynamics. It is really working out to be a dramatic clash of fleets though ... surprisingly quick resolution of the conflict beckons.

Thanks for the offer of an alternative computer to resolve the hosting but I don't really think it is that much a matter of processor speed.

Compare ... default turns (with fighter engagement plotted) 13+ hours (still rising) on a fast NT Server...

turns with one player adjusting his fighter strategy to defensive clumps at home ...

hosting takes 3 mins (more or less as usual)

A somewhat significant difference!

Re: pbem hosting hell!
 

Hehe - I wasn't offering my computer (P133!). I just thought it might be a problem specific to your machine and so it might be worth trying it on a diffrent computer.

Re: pbem hosting hell!
 

Anybody have a couple hours of timeshare available on a Cray?

http://www.shrapnelgames.com/ubb/ima...ons/icon12.gif

Re: pbem hosting hell!
 

Might be a good idea to mail the offending turn to MM and tell Aaron it took forever to process. Hopefully he can do something to speed up the process.

Re: pbem hosting hell!
 

I seriously douubt that it is going to finish, or that it's actually doing anything at this point. 1.35 seems to be better than 1.30 at avoiding the RCE's and app errors, but I have seen a couple games that simply never process. This sounds to me like what you are describing.

If you check the task list it will probably tell you Se4 is not responding.

Geo

Re: pbem hosting hell!
 

ho hum, similar problem with the next turn I tried to process, this time the hosting program pauses at 60% and remains there for a long long time. (1 hour and counting this time)

Shortly before the pause begins the "end turn" text disappears from the main display window - not sure if this is connected or a symptom.

The latest turn in my pbem game is likely to include less fighter combat (since we were all warned off Last time) - so now, the solution workaround is not going to be so easy.

Does anybody know if it is possible to process 1.35 games with a 1.30 host? Would that be likely to help (for the fighter bugs alone)

As always, help and comments welcome

Re: pbem hosting hell!
 

btw ... i have mailed the game turn and player files involved to Aaron ... it's a nasty problem, particularly with a view to play-by-web automated hosting in the future;

How would the players know if the host was working or not, or busy processing a seemingly infinate turn?

Although the pbem game involved is high end with large numbers of ships and fighters (1000, and 20,000) respectively - I can see similar situations occuring with future games.

****************

Re: pbem hosting hell!
 

we have a similar situation in a test game running on the seiv.pbw.cc (Admiral's PBW server) where the game is hanging on a massive fighter battle we set up just for this occasion.

I have to commend the Admiral here, and put out this plug for the SEIV PBW service (runs great, no complaints, lots of Users).. here is the way we have to work arround it on the server:

the server will let you know that the turn has crashed, the designated host of the game is then able to either work it out on his own on his own PC and re-upload a manual turn file, or the offending players can be instructed to back-off so that the turn can be reprocessed. if all else fails, the PBW administrators are always willing to help people troubleshoot and work out any problems.

so I am happy to say that PBW handles this, and hopefully when the new Version comes out, there will be less hangs to deal with.

Re: pbem hosting hell!
 

*happy news* - the hosting program managed to process the latest turn after more than 2.5 hours doing an impression of a crashed application (the dialogues had disappeared, the process counter had gone, the screen had gone white) But lo! Like a returning hope the message eventually popped up and gave us the files...

A right royal pain in the butt these long turns ... but, nice to know that the program can actually continue working to produce a result eventually. I have more faith now that the original problem turn (the 13 hour plus one) would have eventually worked.

Amazing really, the problem elements which caused this one to go on forever look to be some absolutely huge organic seeker ship fleets (100ish base ships, 100ish smaller ships, good proportion armed with seeking parasites, 14-20 per vessel) - that's a lot of seeker heads in a small area of space.

Still, the war goes on....

Regards

Re: pbem hosting hell!
 

I host PBEM games and have had the computer take 20 minutes to process a turn:
11 years into the game, 10 empires left from 17, multiple fleets of 50+ common with a 500 ship per player limit.

The game will seemingly stop for 3 mintues then the progress meter will go up a percent.

I have had the game hang on 2 occassions. These hangs were because of some weird bug with fleets that malfador said they were fixing in the next patch.

Hope that helps